Cocaine has been used by over 30,000,000 Americans since 1980 and frank addiction afflicts at least 1,700,000(1). The medical and social consequences of this stimulant abuse are well known and range from acute psychoses to cardiac failure and from violent behavior to crack-addicted newborns(2-4). Cocaine-induced disinhibition and an increased propensity for high risk behavior now pose a special peril with the advent of the acquired immunodeficiency syndrome (AIDS). The highly reinforcing nature of stimulants makes this form of substance abuse especially pernicious and despite a variety of pharmacologic and non-pharmacologic approaches to treatment, no modality is adequately successful(5,6). The reinforcing potential is clearly related to peak serum concentration(7-9). Also, the rapidity with which the peak is achieved appears critical and may relate to the observation that tolerance to the psychopharmacologic and physiologic effects of cocaine manifests during the course of a single administration(10). rampaging abuse of crack, a smokeable form of cocaine, likely corresponds in part to its rapid delivery across the lung with an efficiency approaching that of an intravenous injection(1,5). Pharmacokinetics may also explain the propensity for binge use associated with crack smoking(1). An agent that reduced the velocity to and magnitude of peak serum levels would permit this hypothesis to be tested as well as have major therapeutic potential.
The neuropharmacologic approach to treatment has focused on receptor systems such as the dopaminergic pathways that mediate the effects of cocaine(11). A direct antagonist to cocaine is not available but agents such as desipramine show some promise for maintaining abstinence(12,13). However, there is a lag of several weeks in the onset of desipramine's effect and during this induction period a marked potential for recidivism remains(5,14). An agent effective even for just this period could have important clinical applications but at present no such agent exists. An alternative to receptor based approaches would be to interfere with the delivery of cocaine to the central nervous system (CNS) so that a dose of cocaine no longer had a reinforcing behavioral effect. Since there is no prospect for excluding cocaine from the circulation, this approach would require binding of cocaine by a circulating agent.
In the 1970's Schuster and colleagues investigated an immunologic approach to substance abuse based on the possibility of interference with CNS delivery(15). A rhesus monkey was allowed to self-administer heroin to dependence, and then was immunized to an opiate. Despite access to the heroin, the animal no longer self-administered it. The serum anti-opiate antibody titer greatly exceeded the cerebrospinal fluid titer and this localized the antibody effect to the serum. Thus, the association of heroin and circulating heroin antibody must have been sufficiently rapid to block the heroin's effect. However, the limitation of the approach was identified in that continued administration of very high doses of heroin exhausted the pool of circulating antibody and the animal resumed heroin self-administration. Thus, the approach worked in that the antibody effectively bound the drug and did modify behavior but the approach was limited in that the antibody supply was exhaustible. An antibody would need the characteristics of an enzyme to avoid being "depleted" itself as it depleted its target.
Recently, the exciting development of catalytic antibodies has been reported(16,17). Catalytic antibodies not only bind but also act as artificial enzymes which metabolize their target thus freeing the antibody for further binding(18-25). The principles of this startling advance are illustrated by considering the hydrolysis of a carboxylic acid ester by an enzyme. As seen in FIG. 1, hydrolysis of the planar ester commonly proceeds through a tetrahedral intermediate which decomposes to yield alcohol and planar carboxylic acid. The rate of the reaction varies with the magnitude of the activation barrier(.DELTA.G) between the starting ester and the peak or transition state structure. An enzyme's active site contains a pocket that complements the structure of the hydrolysis transition-state and through various binding interactions, the enzyme stabilizes the transition-state relative to the starting material. This differential stabilization decreases .DELTA.G and contributes to catalysis. The transition state corresponds to a particular configuration of atoms and is thought to resemble the definable species closest to it in energy, i.e. the tetrahedral intermediate in the case of ester hydrolysis. The transition state is unstable and evanescent but phosphonate monoesters are stable compounds which resemble this species in geometry and distribution of charge and on this basis, may serve as transition state analogs. An antibody elicited to such an analog will manifest binding interactions complementary to the hydrolysis transition state being modeled. This antibody, by binding to the modeled substrate, will stabilize the transition state relative to the starting state, lower the activation barrier and catalyze the hydrolysis. By binding and destroying its target, the catalytic antibody is then freed to bind an additional target. Ample literature precedent supports the use of catalytic antibodies as artificial enzymes for the hydrolysis of esters(16,17,26-33).
Of all the commonly abused substances, cocaine is the best candidate for this approach. Attached to the ecgonine nucleus of cocaine is a benzoyl ester group which when hydrolyzed results in a virtually inactive product(35,36)--this is one of the pathways of deactivating metabolism in humans(35,36). The transition state of that reaction resembles the tetrahedral intermediate of hydrolysis and can be mimicked by a suitably designed phosphonate ester analog of the hydrolysis transition state of the cocaine benzoyl ester. A subpopulation of the antibodies elicited by this cocaine analog will function as esterases highly specific for cocaine. Thus, the principal impediment to the immunologic approach suggested two decades earlier--the exhaustibility of the circulating antibody--could be overcome. The anti-cocaine catalytic antibody generated ##STR1## in this fashion would destroy cocaine and be itself available for continued function. The application of such a reagent antibody to the problem of chronic cocaine abuse would be to deprive the abuser of the reinforcing effect of the drug, thereby providing a window for appropriate psychosocial and relapse prevention interventions, and promoting extinction of the addiction.
Only a subpopulation of anti-analog antibodies will possess catalytic activity; so the production of a monoclonal antibody and passive immunization of subjects is required(37,38). Monoclonal antibodies have become established pharmaceutical agents for the treatment of organ transplant rejection(39) and Gram negative septicemia(40). Passive immunization with an anti-cocaine catalytic monoclonal antibody appears to be practical in clinical terms. Second is the duration of effectiveness. The currently available monoclonal pharmaceuticals are administered daily since, as these antibodies bind, the antibody-antigen complexes are removed from the circulation. In contrast, a monoclonal antibody functioning as an artificial enzyme could be designed for longevity(41)--the Fc portion of the antibody genetically engineered for a low clearance rate and portions of the antibody "humanized" by substitution of human in place of mouse epitopes to reduce antigenicity and clearance by a host immune response(42,43). Ideally an administration of an artificial enzyme against cocaine would last for several weeks and provide the extended coverage important for populations with a record of poor compliance.
Third, the efficacy of an artificial enzyme against cocaine relies on a kinetic argument that the rate of cocaine destruction will be able to match the rate of delivery to the CNS. In order to specify the kinetic requirements for the anti-cocaine catalytic antibody, a kinetic model of cocaine delivery is needed. If a dose of smoked crack is absorbed across the lung over the 90-120 second period of one circulation of the intravascular volume (the volume of distribution for the antibody) then an even mixing of cocaine and antibody pools may be assumed. The volume of distribution of cocaine is over twice the total body water(44), but we may neglect this since partitioning of cocaine into other compartments would only decrease on antibody activity. From the moment of cocaine and antibody mixing in the lung, approximately fifteen to twenty seconds elapse before cocaine reaches the CNS capillaries and the most stringent criterion would require complete destruction of cocaine by that time. For a large 100 mg dose (0.36 mmoles) and complete hydrolysis in 15 seconds, the required rate is 0.023 mmol/sec. Thus, the product of the quantity of enzymatic antibody and the antibody's intrinsic enzymatic turnover rate must exceed this value. The assumptions in this model are uniformly conservative and if liberalized would decrease the demand on enzyme performance accordingly. Thus at a monoclonal dose of 200 mg (the monoclonal HA-1A (40) is dosed at 100 mg) the required turnover-rate would be on the order of 2 sec.sup.-1 to 20 sec.sup.-1. Catalytic antibodies have been reported with esterase turnover rates from 20-40 sec.sup.1 and although these esterases were directed toward particularly susceptible target esters, activity of this order of magnitude is possible(20,29). Also, the K.sub.m values for artificial esterases are as low as 2-15 .mu.M (20,31) less than likely pulmonary venous concentrations of cocaine from crack inhalation. We conclude that the kinetic requirements for a clinically useful anti-cocaine catalytic antibody are attainable. An added advantage is that an antibody suitable for the treatment of addiction by the above criteria could be suitable for the treatment of acute overdose. A final concern is the possibility of saturating the enzyme with massive dosages of cocaine. However, the reinforcing effect of cocaine may not be as significant if peak serum levels are reached more gradually and the large dose of crack may be blunted in effect to a weak dose of nasal cocaine hydrochloride. Thus, the protection afforded by an anti-cocaine enzymatic antibody may not need to be complete in order to be useful.